Abstracts

Rationale: Electric source imaging (ESI) and magnetic source imaging (MSI) are non-invasive tools to aid localization of the epileptogenic zone in epilepsy surgery candidates.  There is large variation between epilepsy surgery centers regarding ESI/MSI availability, clinical indications, hardware specifics and applied mathematical models (Mouthaan et al., 2016). In light of the E-PILEPSY project, which aims to harmonize and optimize presurgical diagnostic procedures across Europe, we carried out a systematic review to assess the diagnostic accuracy of interictal high resolution (HR-) ESI and interictal MSI to localize epileptogenic regions of interest in epilepsy surgery candidates. Methods: Embase, Pubmed and the Cochrane database were last searched on 13 February 2017. E-PILEPSY taskforce members performed article screening, data extraction and quality appraisal. We included studies that focused on diagnostic accuracy and used seizure outcome after surgery (at least 6 months follow up) as a reference standard. Data were extracted with quality appraisal of individual studies based on the QUADAS-2 framework. Sensitivity was defined as the proportion of patients with good-outcome with sources localized within the resection volume compared to the total number of good-outcome patients. Specificity was defined as the proportion of patients with poor-outcome with sources localized outside the resection volume compared to the total number of poor-outcome patients. Sensitivity and specificity values of studies were combined into summary estimates by means of a bivariate linear mixed model.  Results: The electronic search yielded 1964 publications. Eleven studies were included: nine on MSI (n=259), three on HR-ESI (n=127). Selected populations included general surgical candidates and MRI negative patients. A realistic head model was used in one HR-ESI study, and  spherical head models in two other studies. All used linear distributed inverse solution. Equivalent current dipole (ECD) was used in all MSI studies, however cluster definition varied. Study quality varied between studies, although no study was completely free from bias. Bias mostly involved the following aspects: 1- inclusion restricted to patients who underwent surgery (disease spectrum bias) 2- decisions to proceed to surgery/demarcation of resected area (partially) guided by MSI or ESI (partial verification bias), 3- exclusion of indeterminate test results (e.g. insufficient number of IEDs). Summary sensitivity and specificity were 71.5% (95% CI:61.6-79.7%) and 60.3% (41.4-76.6%) for MSI, and 85.6% (76.7-91.4%) and 67.7% (49.3-81.9%) for HR-ESI. Conclusions: High quality evidence to support the added value of MSI and ESI in localizing epileptogenic tissue in epilepsy surgery candidates is lacking. All studies are highly biased on numerous aspects and some varied regarding the included population and test methodology. High quality studies, allowing unbiased test evaluation, are needed for evidence-based recommendations regarding the application of MSI and ESI in the presurgical workup of refractory focal epilepsy. Funding: This work arose from the project E-PILEPSY, which has received funding from the European Union, in the framework of the Health Program (2008-2013).